Going to Surprise: the grammaticalization of itive as mirative Daniel Ross University of Illinois at Urbana-Champaign [email protected] — danielrosslinguist.com Abstract Morphemes indicating direction away from the deictic center can signal an unexpected event, without necessarily indicating motion in space: Look what he went and did this time! These surprise readings have puzzled researchers working on several European languages, especially English and Swedish (Stefanowitsch 1999; Josefsson 2014). In this paper it is proposed that mirative usage like this can be explained as metaphorical movement away from an expected outcome: the way an individual or the world is supposed to behave according to the speaker. Just as motion verbs can grammaticalize from the spatial domain to the temporal domain as aspect markers, they can also extend into the domain of eventualities. Specifically, go and related forms can grammaticalize as miratives, marking unexpected information (DeLancey 1997). It is first important to recognize a category of deictic modification for verbs. Various forms are attested cross-linguistically, from verb-verb combinations (such as serial verb constructions) to inflectional morphology in some languages. These morphemes may either add a motion sub-event to the verbal predicate (Associated Motion: Guillaume 2016), or encode only the direction of an existing motion event (Directionals). In both cases, the two most common types are action directed toward the deictic center (usually the speaker) and action directed away from the deictic center. The former corresponds to come and is known variously as ventive, venitive, cislocative and centripetal. The latter, corresponding to go, is the focus of this paper, and is known variously as itive, andative, translocative and centrifugal. In a large sample of languages with verbal deixis of this type, although not as common as grammaticalization into tense and aspect markers, itives of various forms were found to express mirativity, from Abkhaz (Caucasian) to Kera (Chadic) to Mizo (Sino-Tibetan) to Turkish, supporting an analysis of these constructions as a unified category. Semantically, several factors play into this grammaticalization pattern. One is that go is often associated with negative actions (go crazy), and come with positive ones (come to your senses). Another is that go often carries a sense of intentionality or inchoativity even in tense and aspect (Fleischman 1982). Sudden or surprising action is encoded by the speaker as deviation from an expected course of action. Interestingly, though less often, expected or desirable outcomes can be expressed by ventives as, for example, hortatives or solidarity imperatives (semantically like English c’mon; cf. Heine & Kuteva 2002). Online proceedings of Cognitive Linguistics in Wrocław Web Conference 2016 https://sites.google.com/site/coglingwroc2/ “Somewhere along the way I lost a step. I got sloppy, dulled my own edge. Maybe I went and did the worst crime of all: I got civilized. So now we zero the clock, just me and this no-name world. Gotta find that animal side again.” 1. Introduction In the opening scenes of the 2013 science fiction film Riddick, the anti-hero, who is the last person anyone would imagine as civilized, describes what went so wrong in his life that he would end up abandoned on an uninhabited planet. This use of the motion verb go in English can indicate what has been called a “surprise” reading: (1) Look at what he went and did this time! This construction thus joins the other better-known functions of go, including its use as the motion component of multi-verb constructions as in (2), its metaphorical extension to so- called fictive motion (Talmy 1996; Matlock & Bergmann 2015; inter alia) as in (3), and its grammaticalization in the temporal domain as in (4): (2) We went to see a movie. (physical motion) (3) The road goes north. (fictive motion) (4) We’re gonna see a movie. (future) Compared to these other constructions, research on the surprise function of go is relatively sparse and mostly limited to European languages, especially English and Swedish. Most of the relevant discussions have been limited to pseudocoordination constructions, which have been most extensively described and investigated in European languages (cf. Ross 2016a). It should be noted that while pragmatic effects can make one reading especially salient, as in (1), English go and can be interpreted either as literal motion or with the surprise reading, as in (5): (5) He went and bought that book. In many cases the surprise reading of a pseudocoordination construction and the construction itself are considered idiomatic, and previous accounts have rarely related this to other uses of GO cross-linguistically, and not at all to morphological itives (see Section 2). That is therefore the goal of this paper, both to establish a general cross-linguistic grammaticalization pattern and to shed more light on the properties of these European pseudocoordination constructions. The remainder of this section provides an overview of previous research before turning to cross-linguistic comparisons and a proposed explanation. Ross (1967:170), in identifying uses of and that are not subject to certain structural properties of coordination (specifically the Coordinate Structure Constraint, which prohibits asymmetrical extraction from conjuncts), points out (6) as anomalous and beyond the scope of his research (because it is not to be derived from structural coordination). He does not discuss the surprise reading explicitly, but the example is rather transparent: (6) Which dress has she gone and ruined now? In discussing pseudocoordination (especially the types try and do, and go and get; as well as go get), Carden & Pesetsky (1977:89) identify an “unexpected event” reading for go and. They are careful to distinguish unexpected event go and from motion go and because, although they claim both are not derived from structural coordination and thus involve a 2 “fake” and, the unexpected event usage usually has a negative connotation, and it does not require that the subject be in control of the event, as in (7): (7) Did your best coon dog go and die on you?1 Arnaiz & Camacho (1999) discuss a similar construction in Spanish. They claim that ir y ‘go and’ functions as a Topic Auxiliary, and the focus of their analysis is on its syntactic structure from a Minimalist perspective. As in English, the construction describes an action that “suddenly alters the expected course of events” (p.319): (8) Ramón fue y se cayó. 2 Ramon went and REFL fell ‘Ramon unexpectedly fell.’ (Spanish: Arnaiz & Camacho 1999:318) Stefanowitsch (1999, 2000) provided the first cross-linguistic investigation of go and, with examples from English, Swedish, Danish (9), Spanish and even Hebrew (10): (9) Han var gået hen og have gifted sig. He had gone there and had married REFL ‘He went and got married.’ (Danish: Stefanowitsch 1999:126) (10) Kulam paxadu liftoax et ha-kufsa aval Dan halax ve asa et ze. everyone was.afraid to.open OBJ DEF-box but Dan went and did OBJ it. ‘Everyone was afraid to open the box, but Dan just did it.’ (Hebrew: p.127) Stefanowitsch (1999) was also the first to attempt an explanation of the surprise reading from the semantics of the verb GO, extending ideas from Ekberg (1993) about ingressive use of TAKE. He observes that GO involves a motion schema which may take several variant forms including one involving divergence from an expected direction. This is extended through the metaphor CHANGE IS MOTION (see Radden 1996), and divergence from the expected flow of events is surprising. This analysis, to be revised and discussed in more detail, is the starting point of my analysis in this paper (Section 4). Several studies have investigated the surprise reading of GO in Swedish. Wiklund (2008, 2009) describes the inceptive use of gå och ‘go and’ (as well as ta och ‘take and’): (11) Peter gick och läste en bok. Peter went and read a book ‘Peter went and read a book.’ (Swedish: Wiklund 2009:181) She claims that surprise is built on top of the inceptive reading, arguing that punctuality is critical for deriving surprise. For GO, the second, non-motion verb contributes punctuality, while TAKE can contribute punctuality directly as the first verb. Her analysis is formal, decomposing event structure, and it adequately describes the distribution of surprise readings but does not explain why motion verbs are used as light verbs. Josefsson (2014) also describes Swedish gå och but offers a very different explanation. She relies on the polysemy of GO and bases the surprise effect on the ‘happen’ 1 Their particularly colloquial example suggests the authors recognize that this usage is typically informal. 2 Glosses are presented for the most part in the same format as in the original source, but in some cases I have adjusted the glossing to clarify the relevant phenomenon or supplied the glosses myself when the author did not. A list of abbreviations used in the glosses is included at the end of the paper. 3 meaning. Specifically, she claims that GOHAPPEN has three thematic roles (12) but that the light verb in ‘go and’ has only two argument positions so that EXPERIENCER cannot be expressed structurally; the speaker assumes this role, experiencing the event as a surprise. (12) It went well for him. (GOHAPPEN thematic roles) THEME GOAL EXPERIENCER Although this explanation is interesting, it relies on the tentative and anomalous theoretical assumption that an interlocutor can take on a grammatical role in a sentence. I propose an alternative explanation in Section 4 that does not rely on such an assumption, and it also follows naturally from the primary motion sense of GO. As mentioned by Stefanowitsch, the motion need not be literal, but may be interpreted as a change of state.